Subject: ESCApad #14 Date: Mon, 31 May 1999 13:18:01 +0100 From: Isabel Trancoso To: esca_list@speech.inesc.pt ============================================================================ ESCApad number 14 May 31, 1999 ============================================================================ Dear ESCA members, Welcome to ESCApad #14. This issue has several announcements including two Special Issues and two PhD Theses. Greetings from ESCA, Isabel Trancoso ======================================================================== PAPERS ON LINE: >From the ESCA Tutorial and Research Workshop which was held on the 19th and 20th April 1999 in Cambridge, UK: http://svr-www.eng.cam.ac.uk/~ajr/esca99.html ======================================================================== Look up Speech Central, a new directory source for speech recognition and related technology information, including business, research and educational links: http://www.speechcentral.com/ ======================================================================== Pre-defense of two PhD theses at ENST (Paris, May 1999): 2D statistical analysis for segmental modelisation of speech signal: Application to recognition tasks. Guillaume Gravier http://www.doctorat.enst.fr:8080/doctorat/THESES/Guillaume.Gravier.html Contribution to the verification of a person's identity by using data fusion Patrick Verlinde http://www.doctorat.enst.fr:8080/doctorat/THESES/Patrick.Verlinde.html ======================================================================== SPEECH COMMUNICATION CALL FOR PAPERS Special Issue on SPEECH ANNOTATION AND CORPUS TOOLS [www.ldc.upenn.edu/annotation/specom.html] Guest editors: Steven Bird and Jonathan Harrington Scope of the Special Issue Submissions are invited for a special issue of Speech Communication on Speech Annotation and Corpus Tools. The aim of the special issue is to make speech scientists aware of recent developments in the representation and management of annotated speech corpora, i.e. collections of speech signal data with time-aligned transcriptions. (Signal data may be audio or physiological, natural or artificial, in basic or derived form.) The primary focus is the structure of annotations and of annotated corpora, as used within and across a wide range of disciplines concerned with spoken human communication. Annotated speech corpora have been a critical component of research in the speech sciences for some years. Today, these corpora are being created and deployed for a rapidly expanding set of languages, disciplines and technologies. A wealth of formats and tools have sprung up around this enterprise, a diversity which at once facilitates and frustrates progress. The linguistic annotation page [www.ldc.upenn.edu/annotation/] has drawn attention to the scale of ongoing activity, to the existence of diverse approaches to similar problems and of similar approaches to diverse problems. Despite the explicit formats and well-documented user interfaces, insights about the structure of the annotations themselves are often buried in coding manuals and internal data structures. There is a pressing need for papers which document the corpora and tools, which identify notational and functional equivalences among different approaches, and which report on new approaches to core representational problems. The special issue will consider papers which address theoretical and practical issues concerning the representation of annotations, the structure of annotated corpora, and the design, analysis and implementation of tools for creating, browsing, searching, manipulating and transforming annotations and annotated corpora. In each case, the description of annotation structures or tools should be accessible to readers outside the particular community in which the system originated. A broad sampling of relevant issues is given below: + representational issues: - sequence, overlap, hierarchy - simultaneous cross-cutting hierarchies - the nature of labels - pointers and cross-references - temporal structure, instants and periods - atemporal information (e.g. demographic data) + relationships between annotations and signals: - multiple independent annotations of a single signal - single annotations which reference multiple signals - annotations which reference other annotations + database issues: - structuring annotations, signals and atemporal data into a corpus - indexing for efficient access of large corpora - high and low level query languages, cross-level query - validation, update, provenance - data transformation and integration - file formats, storage, transfer; the place of XML + implementation issues: - design philosophies and functionalities for annotation toolkits - approaches to creation, browsing, navigation, display - reusability, interoperability, platform independence - integration with independent tools (e.g. statistical analysis) - techniques for working with multiple corpus formats + wider issues: - methodologies for research and development involving annotated corpora - the cycle of refining annotations and refining theoretical models - the role of annotated corpora in evaluating theories and systems - necessary steps towards general purpose tools and formats Important Dates * 400 Word Abstracts: any time in May-July * Advance Notification: Monday August 16th, 1999 * Submission Deadline: Monday August 30th, 1999 * Acceptance Decision: late October, 1999 * Final Version Due: late January, 2000 * Publication Date: mid 2000 Advance Notifications 1. Prospective authors are encouraged to submit a 400 word abstract of their paper so that the editors can comment on its suitability for the special issue. These abstracts should be formatted as ASCII text and submitted by email to both editors. 2. To facilitate a rapid review process, authors are required to give notification of their submission two weeks in advance of the submission deadline. Notification should consist of the title and (a draft of) the final abstract, formatted as ASCII and emailed to both editors. Submissions All submissions must consist of original unpublished work that is not being submitted for publication elsewhere. Papers should be approximately 30 pages double spaced. Electronic submission is encouraged. Details about preparation of electronic and paper submissions, and any updates to the CFP, will be posted on the web at [www.ldc.upenn.edu/annotation/specom.html]. Please register at this site to receive email notification of any subsequent announcements concerning the special issue. - - -- Dr Jonathan Harrington Director, Speech Hearing and Language Research Centre, Department of Linguistics, Macquarie University Sydney, NSW 2109, Australia. Tel: +61 2 9850-8740 Fax: +61 2 9850-9199 jmh@srsuna.shlrc.mq.edu.au http://www.ling.mq.edu.au/dbase/person.phtml?oid=19084 Dr Steven Bird Associate Director, Linguistic Data Consortium University of Pennsylvania, 3615 Market St, Suite 200 Philadelphia, PA 19104-2608, USA Tel: +1 215 573-3352 Fax: +1 215 573-2175 sb@ldc.upenn.edu http://www.ldc.upenn.edu/sb ======================================================================== CALL FOR PAPERS Natural Language Engineering Special Issue on Best Practice in Spoken Language Dialogue Systems Engineering NLE SPECIAL ISSUE AS A DISC INITIATIVE A special issue on Best Practice in Spoken Language Dialogue Systems Engineering will be published by the journal of Natural Language Engineering (NLE; Cambridge University Press) in the beginning of 2000. This issue is an initiative of the European Esprit project DISC (June 1997-December 1999), formally called "DISC Spoken Language Dialogue Systems and Components. Best practice in development and evaluation". The main goal of DISC is to identify current practice in the development and evaluation of Spoken Language Dialogue Systems (SLDSs) and their components, in order to come to a definition of best practice. DISC intends to contribute to the establishment of dialogue engineering guidelines to be used by different target groups, among others developers, deployers and customers. BACKGROUND AND OBJECTIVES The interest in SLDSs has increased enormously over the last few years: at present there is a large number of systems available many of them for commercial use; the number is growing rapidly and so is the variety of functionalities and domains of application. These developments have led to a situation in which there is a great need, shared by developers, deployers and customers alike, for effective guidelines, which will enable them to make well-formed design and implementation decisions, in accordance with broad consensus of what must be 'best practice' in this particular engineering domain. The purpose of this special issue is to bring together leading views on what might be considered to be best practice in the development and evaluation of SLDSs. We are aware that this is a delicate notion - what constitutes best practice depends on the kinds and complexity of tasks the SLDSs are to perform (e.g., with increasing task complexity, the need for improved dialogue control requires more sophisticated control of input speech and input language processing) and on a number of other constraints on SLDS development, having to do with resources available for system development, the constraints imposed by the different groups involved (e.g., developers' constraints, customer preferences and user group defined constraints), etc. So, we would like to take as a starting point a definition of best practice relative to factual constraints imposed on SLDS development. THEME In agreement with the main goal of DISC, the general theme for the special issue is what could be taken as best practice in SLDS engineering, given the availability of different technological options with their inherent merits and limitations which are subject to different constraints on system (component) realization. We are interested in new, high quality papers which address, along the lines of the objectives above, one or more of the following issues: (i) best practice in the development and evaluation of SLDSs as a whole or (ii) best practice in the development and evaluation of one or more of the following system aspects, as well as of the interaction between them: - speech recognition - speech synthesis - natural language understanding and generation - dialogue management - human factors - system integration All papers should fall within the scope of NLE, as described in the instructions for contributors to the journal. This mainly implies that the research views, comparative discussions, etc. described in the papers must have a clear potential for practical application, in this particular case meaning that they contribute to guidelines for SLDSs best practice (see also the NLE web page, the reference of which is given below). SUBMISSIONS Submissions to the special issue should be in line with the NLE style sheet, which is obtainable via the NLE web page. The length of a paper should be 10-12 journal pages. Electronic submissions should be sent as a postscript file by e-mail to the co-ordinating special issue editor. Alternatively, 6 hardcopies can be sent to the editorial address given below. The deadline for submission is September 1, 1999. Authors are asked to e-mail a short statement of their intention to submit a paper to the co-ordinating special issue editor before July 15, 1999. REVIEW PROCEDURE All papers, both those submitted by members of DISC and from outside the project, will be double reviewed and triple reviewed if necessary. The review committee consists of seven members of the DISC consortium, one member of the DISC Advisory Panel, three members of the NLE editorial board and a group of ten external referees. In case of a very large number of submissions the review committee will be extended accordingly (see website). IMPORTANT DATES - - - Intention to Submit Due Date: 15 July, 1999 - - - Paper Due Date: September 1, 1999 - - - Revision Due Date: December 15, 1999 - - - Acceptance Date: January 2000 - - - Publication Date: February/March 2000 SPECIAL ISSUE EDITORS The special issue editors are the IMS group participating in the DISC project: Jan van Kuppevelt (co-ordinating editor) kuppevelt@ims.uni-stuttgart.de Ulrich Heid heid@ims.uni-stuttgart.de Hans Kamp kamp@ims.uni-stuttgart.de Editorial Address: NLE Special Issue c/o Jan van Kuppevelt Institute for Computational Linguistics (IMS) Azenbergstrasse 12 D-70174 Stuttgart Germany Tel.: +49 711 1211357 or 6574548 Fax: +49 711 1211366 FURTHER INFORMATION Web site for Special Issue: http://www.ims.uni-stuttgart.de/NLE_special_issue/ (under construction) Web site for NLE: http://www.cup.org/journals/jnlscat/nle/nle.html Web site for DISC: http://www.elsnet.org/disc/ ============================================================================= All additional information at the ESCA web-site: http://www.esca-speech.org The ESCA secretariat can be contacted at: info@esca-speech.org Requests concerning membership, Speech Communication and ordering Proceedings should be forwarded to the Secretariat. For message distribution at esca_list contact: publicity@esca-speech.org =============================================================================